Laser scanning at Castle Hochosterwitz in Carinthia.

Transcription

14th International Congress „Cultural Heritage and New Technologies“ Vienna, 2009
Laser scanning at Castle Hochosterwitz in Carinthia.
3d optical documentation techniques in the service of a time-efficient, highly
precise and distortion-free architectural survey
1
Gerold EßER / Jan KANNGIEßER / Mathias GANSPÖCK
1
Vienna University of Technology, Department of History of Architecture and Building Archaeology
Abstract: The paper delineates a strategy capable of creating a number of efficient and practical outcomes
that can be generated from 3d scanning data. In the field of architectural heritage conservation especially,
and in all cases where substantial transformations have to be implemented, pure 3d outcomes do not really
meet the needs of planners. While on the one hand a full surface 3d recording of the artefact is welcome to
the national preservation authorities for documentation needs, planning architects are still forced to deal with
2d materials.
The strategy developed for the laser scanning campaign at Castle Hochosterwitz shows that both demands
can be satisfied without the need of additional tachymetrical measurements. However density and
geometrical precision of point clouds make sure, that 2d planning materials up to a certain scale can be
deduced without loss of exactness and reliability. The innovative new workflow established has also shown
to be more time and cost efficient, when compared with conventional documentation procedures. It is also
more flexible: once a complete and homogeneous data set of the object containing 3d geometry and texture
has been recorded, all kinds of outcomes – ground plans, sections, top views, prospects, ortho-photos - can
be derived without the need to go back on site.
Keywords: Laser scanning – architectural survey – documentation strategy – optimized workflow
Introduction
Laser scanning is becoming a more and more refined and powerful working tool in cultural heritage
documentation. In the field of archaeological and architectural documentation, and especially when
combined with digital photogrammetry, it reveals a stunningly high potential which finally will question wellestablished documentation procedures based on a purely tachymetrical approach. Compared with
conventional documentation techniques it strikes by its obvious capacity in documenting complete 3d
surfaces thereby creating a consistent and textured 3d image of the historical artefact (ZIMMEMRMANN and
ESSER 2008). All kinds of 3d modelling approaches have therefore been introduced into archaeological
research projects in spite of often being time-consuming and difficult to handle. It might be because of the
substantial problems inherent in the elaboration of 3d data that the scientific efforts concerning 3d matters
seem to have covered relevant parts of our intellectual capacities. In consequence more easy and practical
approaches which focus on reducing large quantities of information recorded in 3d scanning campaigns have
rarely been adopted in order to acquire conventional but easy-to-handle planning materials. The problem
seems to have a psychological dimension too: Do we consider research dealing with 3d problems more
worthwhile than caring about the evident necessity of building a bridge between 3d research and recording
309
and 2d usage of geometric data which still represents the standard in the world of planning? Does it seem
anachronistic to use a full 3d data set but focus on a final 2d output? The evident void between an
innovation-oriented research field of 3d measurement technologies and a performance-oriented approach in
the world of planning and building has soon to be bridged. It is time to think anew about 2d without falling
back into a practice that seems to be outdated.
Fig. 1 – Castle Hochosterwitz, north-west view (Copyright: Jan Kanngießer)
There are quite a number of good arguments why a combination of the two approaches seems reasonable.
Firstly, documentation in 3d space has the power to record the status of an historical object in an allembracing way, by measuring its full surface geometry and by recording real colour textures for all the
measured object surfaces (ESSER and MAYER 2008). The result is a 1:1 image of the artefact which can be
preserved as a precious documentation of its specific status at a certain moment in time. Secondly, 3d
scanning technologies have gained a stunningly high velocity and productivity in measuring complex 3d
spaces; without doubt, they have the power to compete with conventional tachymetric measuring strategies.
Not only has it become evident that the time necessary for positioning and orienting a laser scanner is even
shorter when compared to a tachymetric system and that the precision of geometric data recorded by laser
scanners can already be compared to that of total stations. But more importantly, the amount of information
and richness in detail of any laser scanning data coverage easily exceeds the information recordable by the
most experienced operator of tachymeters. When combined with digital photography the two data sets –
geometry and texture – form a hybrid but still homogenous and uniform documentation output that cannot be
gained by a conventional combination of tachymetry and photogrammetry (JANSA et al. 2004; ULLRICH
310
2004). The documentation results obtained by laser scanning devices thus have an unquestionable value
that should not be disclaimed by hesitant or conservative approaches.
A reduction of the complete 3d and texture data into a practical and application-oriented 2d extract thus will
have to be applied with respect to the right point in time in the course of a documentation procedure: once
the 3d recording has been completed and a comprehensive 3d image has been created the possibilities
regarding subsequent works are up in the air again. The elaboration of 2d documentation materials has a
long and successful tradition in architecture. Starting with Renaissance artists like Raffael, Serlio and many
more who were eager to document the standing remains of the Roman past, the 2d representation finally
culminates in the powerful working tool of our times, which – in the best sense of a refined level-of-detail
strategy – is able to build a bridge from the smallest detail representation up to an urban scale. At the same
time however, architectural plans should not be understood as mere reductions of a seemingly more
complex 3d environment. Quite to the contrary they are calculated abstractions that follow a well-established
coding system, an international language able to express all the specific information necessary to describe
the general disposition and all relevant details of a 3d architectural space. Thus architectural plans are in no
way mere reductions but to the contrary they represent rich abstractions and meaningful concretions full of
information which are worth being considered as valuable outcomes of an innovative 3d recording.
Fig. 2 – Earliest representation of Castle Hochosterwitz dating from 1575 (Copyright: Gerold Eßer)
Documentation history of Castle Hochosterwitz
Castle Hochosterwitz is situated in the district of St. Veit an der Glan, about 20 km north of Carinthia´s
Federal Land`s capital Klagenfurt. As for its formidable position on top of a more than 175 meter high solitary
crag and because of its extraordinary state of conservation it attracts about 100.000 visitors every year, a
fact that has significantly contributed in transfiguring the organism into something like a cultural monument of
a national dimension. Once a powerful medieval stronghold, it was transformed into a Renaissance castle on
311
a rock during the 16th century. Inhabited by a powerful family of regional – a little later supra-regional –
importance and finally of European format, it has at all times been the impressive object of historical
representations, executed in frescoes, oil paintings, etchings and drawings (DINKLAGE 1980), the earliest
known dating from 1575. All of these images surely had the contractors´ wish for representation of earthly
power being transformed into an image, which, technically speaking, had to deal with being forced to
condense a complex 3d structure – the castle and its fourteen gates on top of a rock – onto the bidimensional surface of the canvas, a task which has seen great success as all these images bravely
succeed in transporting a faithful view of the structure as a whole and yet in detail, that way putting today`s
historians in a position to be able to reconstruct parts of the history of Castle Hochosterwitz, which otherwise
would not have been documented.
Fig. 3 – Scientific documentation of Castle Hochosterwitz by Paul Grueber in 1925 (Copyright: Verlag Kollitsch, Klagenfurt)
th
The scientific interest in the Castle`s history, that seemed to arise only at the end of the 19 and beginning of
th
the 20 century, however substantially altered the view of an adequate way to represent the building´s
structure. In 1889 and again in 1925 two different maps of the area were published which, with a good
portion of intuition, succeeded in representing the basic conformation of the marvelous ensemble, yet failed
to introduce known surveying standards of the time in order to get distortion-free and dimensionally accurate
layout drawings (CENTRAL-COMMISSION 1889; GRUEBER 1925). The latter author, Paul Grueber in 1925,
when treating constructions of limited dimensions, such as are the famous fourteen castle gates of
Hochosterwitz, clearly showed a great capacity for transferring three-dimensional constructions into
significant and accurate architectural plans, representing each of the gates in floor plans, vertical sections,
prospects, details and perspective drawings with the obvious aim of conveying a comprehensive and
complete idea not only of the geometric aspects but also of the constructional and functional contents of the
displayed objects.
Hence, in looking back on the existing historical documentation material, it seems fair to judge that most of
the representations conserved followed a definite and precise goal and achieved its objective fairly well.
Efficient solutions for clearly understood problems have surely been found which, in most cases, had been
312
developed in accordance with the technological and psychological conditions of that era. But this judgment is
in no way meant to glorify without criticism the representation techniques which are no longer up-to-date. To
the contrary the displayed analysis has been executed with the clear goal of being able to act on a firm and
conscious basis, which will assure that the still inherent, enormous potential for improvement is with
confidence being exploited.
Project and documentation requirements
When in the spring of 2009 the Department of History of Architecture and Building Archaeology of Vienna
University of Technology
75
was asked to undertake an architectural survey of the structures of the highest
level of Castle Hochosterwitz – including all the buildings inside the historical outer ward – it soon became
clear that the planned documentation would, with all probability, constitute a good opportunity to test state-ofthe-art measurement techniques when confronted by a structure that, according to its largeness, complexity
and topographical position, without any doubt belongs to the most challenging objects of its kind. Committed
by Cultural Heritage Authorities and the Government of Carinthia, the documentation was aimed at fulfilling
some quite different purposes. First of all a complete documentation as a faithful testimony to the castle`s
state of conservation seemed necessary; secondly all kinds of yet undefined documentation material was
needed in order to enable a profound and complete historical research of the castle`s history; thirdly a set of
three floor plans and four transversal vertical sections in scale 1:100 was required in order to be able to
conduct the intended restoration process. With respect to scale, quality and detail of information, the
documentation was therefore required to cover the whole organism while still recording the architectural
detail, thus bridging the gap between the scale of a layout drawing and the information inherent in 1:50
plans. Moreover the covered area of about 4.000 square meters, more than 60 rooms distributed over the
three storeys of the castle and the hilly topography of the mountain top, represented conditions that would
have aggravated every kind of conventional surveying strategy strongly. Summarizing all documentation
requirements, it seemed quite unlikely that a tachymetrical survey would – with respect to limited financial
and time resources – be able to satisfy the expectations.
Documentation strategy
It was therefore decided that in principle all utilizable measurements were to be recorded using a laser
scanner and polygonal traverses where measured, in order to construct a rigid net of geo-referenced
reflecting targets as tie-points in order to optimize the registration of the single scan positions. Strong
arguments for that strategy were:
-
Laser scanning would allow the recording a full surface geometry thus enabling the exploitation of
the documentation in the course of further unknown purposes.
-
Laser scanning would guarantee fast on-site progression of the documentation work thus decreasing
personnel costs.
75
For further information see Department homepage http://baugeschichte.tuwien.ac.at
313
-
Additional close-range photogrammetry executed by way of a digital camera mounted on top of the
scanner would allow the recording of real colours as an additive data set.
In comparison to a tachymetrical survey the challenges that had to be faced were to be faster and spend
less time on site, to be as precise as when using tachymetry by working with the highest possible accuracy
of measurement and to record more information thus creating a versatile and flexible data set.
Fig. 4 – Coloured point cloud cluster of St. Nikolaus chapel at Hochosterwitz (Copyright: J. Kanngießer – M. Ganspöck)
For all laser measurements a long-range 3d scanner RIEGL LMS-Z420i was applied.
76
The scanner is
worked on the time-of-flight measurement principle and is operated by RIEGL-own software RiSCAN PRO.
77
At Hochosterwitz a standard angle measurement resolution of 0,12° was implemented creating about
2.000.000 points in space per single scan. This means that in a distance of 5 meters the measured 3d points
form a point grid of approximately 10 mm distance [≈ 5.tan(0,12)] on the object surface. While scanning, only
in the very few cases of extraordinary narrow spaces which had to be documented, this resolution has been
lowered, while in the case of the outer walls of the castle, where greater measurement ranges up to about 20
meters were unavoidable, the resolution was augmented by a horizontal and vertical step width of 0,03° [≈
atan(0,01/20)]. In order to calculate the effect of the standard measuring error on the documentation material
the following considerations needed to be made: the scanner´s measurement error is defined as +/-10mm
standard deviation of the scanning unit measured in the beam direction. With respect to the desired scale of
the documentation plans, a measurement error of 10 mm transferred onto a 1:100 plan would cause a
drawing error of only 0.1 mm; assigned to a 1:50 detailing, the error would be 0.2 mm maximum, still being of
little importance for the accuracy of the drawings. In cases where continuous and quite plain surfaces are
measured the effective accurateness of a plan is even higher as an average determination of a plain wall
surface would with all probability represent the actual position of that wall; in the cases of geometrically more
complex or quite uneven geometries the accurateness of the final plans would on the other hand be
76
For technical specifications see data sheet of LMS-Z420i on RIEGL website http://www.riegl.com/products/terrestrial-
scanning/produktdetail/product/scanner/4/.
77
Technical specifications of RIEGL software RiSCAN PRO are found under:
http://www.riegl.com/uploads/tx_pxpriegldownloads/11_DataSheet_RiSCAN-PRO_22-09-2010_04.pdf
314
considerably lower thus requiring additional measurements using measuring technologies of higher
accuracy.
Special care was taken to maintain the measurement error of +/-10mm when considering the whole
structure. A mere registration of the single scans by extracting local coordinates of all visible reflecting
targets and orienting the scan positions in the manner of a chain registration, would have caused severe
problems. Instead, by introducing a tachymetrical measurement of all the reflecting targets necessary for the
registration of the single point clouds, the overall measurement error that way has been reduced to a
minimum. In addition, all the 3d coordinates of the reflecting targets have been geo-referenced by measuring
three triangulation points found in the vicinities of the castle belonging to the public monitoring network and
by calculating a rotation-translation matrix for the reflecting targets. The 3d coordinates of the reflectors were
then introduced into the RiSCAN PRO scanning project as a new global coordinate system and all the single
scans were subsequently registered on that rigid and geo-referenced coordinate net. By doing so, a high
accuracy of the global registration of all single point clouds was obtained.
During the short period of only five days at the beginning of April 2009, about 250 scan positions were
measured by a team of three with the total result of some 500.000.000 points being recorded. In addition, for
each scan position seven digital photos were automatically captured using a calibrated CANON EOS 1ds
with a 14mm lens mounted on top of the scanning device. Each picture contained about 10 mega pixels of
texture data. Via known internal and external camera calibrations these shades of colour have automatically
been referenced to the geometry of their respective point clouds thus giving the opportunity to create
coloured point clouds of all possible situations required in future evaluation projects. The total number of
calibrated photos recorded is approximately 1750.
By this documentation strategy a coverage of about 95 % of the structure´s surface has been reached,
merely omitting parts of the outer roof surfaces that, due to the fact that they are not visible from the ground,
have not been measured.
Fig. 5 – Scan positions necessary for recording ground plan information; overall point cloud (Copyright: J. Kanngießer – M. Ganspöck)
315
Data processing strategy
The post-processing strategy which will finally lead the way into drawing classical 2d architectural survey
plans, requires some preparatory steps. After the final registration of all single point clouds is finished – a
work that preferably should take place while still on site in order to be able to remedy possible
documentation errors – the scanning raw data sets have to be cleaned of irrelevant information like
vegetation, human beings, furniture et cetera. After that, an important decision has to be taken concerning
the question of whether all single point clouds are going to be reduced to hold a homogenous resolution (e.g.
by applying an octree algorithm or some other comparable filter for point reduction). Note that it is strongly
recommended that the raw data be not reduced but all single point clouds be used with their full information,
to retain the information being evaluated as close as possible to reality.
As a first step let us now imagine the method that will finally enable us to draw a ground plan of a specific
area. All the following steps are still going to be made inside the RIEGL capturing and processing software
RiSCAN PRO. The procedure follows an easy concept: to be able to draw a floor plan all relevant measuring
data-sets have to be selected and exported into CAD (Computer Aided Design) software. A floor plan should
basically contain information about three different, horizontally organized slices of the architectural space: a
section slice which is preferably defined at approximately one meter above floor level; secondly, a floor slice
which contains all the information which is visible below the height of the section slice, and thirdly, a ceiling –
or vaulting – slice which shows all the constructional elements visible above the height of the section slice.
78
As most of the single scans contain information covering more than one of the abovementioned slices, it is
recommended that all relevant single point clouds concerning a given area of the building into one newly
created view are imported. Once this compilation has been completed, the three horizontal slices which are
part of the point cloud cluster are going to be selected, each in one operation, by defining the medium height
of the slice – defined by its absolute z-coordinate – and its thickness. Each slice is saved as a polydata – a
data unit which is a specifically assembled new point cloud – inside the RiSCAN PRO project, exported
using an adequate file format (ascii xyz for the whole point cloud, dxf for the various sections) and finally
opened in a CAD file. As all the recorded point data have initially been registered in one common coordinate
system – in Austria the Gauss-Krüger Coordinate System – and because the XYZ Coordinates of the
polydata are recorded keeping all the original coordinate values, the point coordinates once imported into the
CAD drawing have not been altered.
As previously mentioned, horizontal intersections can be arranged at predefined heights and there is no
limitation to doing so when evaluating a given point cloud. Wherever floor surfaces are heavily inclined or
simply uneven a suitable procedure consists of evaluating the relevant parts by a number of intersections of
equal distance (ESSER 2006a).
78
For further information concerning national and international requirements for elaborating plan material for architectural
documentations consult national norms and rules as e.g. ÖNORM A6250 “Bauaufnahmezeichnungen” for Austria and DIN 1356-6 Teil 6
“Bauaufnahmezeichnungen” for Germany.
316
Fig. 6 – Elaboration of a ground plan: horizontal sections; triangulated mesh; final plan (Copyright: J. Kanngießer – M. Ganspöck)
An alternative procedure can be conducted in all cases when continuously bent or fairly uneven surfaces are
to be interpreted, as for instance vaultings or unsteady floor surfaces showing fairly different height values. In
order to be able to recognize orthogonal projections of section lines of intersecting vaulting surfaces or other
break lines hard to be read from pure point cloud clusters clearly, it is recommended a relevant part of the
point cloud be exported and opened in some kind of 3d modeling software (like GEOMAGIC or
POLYWORKS) and this geometry be converted into a triangulated mesh. The mesh subsequently is to be
shaded for a better reading, transformed into an ortho-photo – a 2d orthogonal projection normal to a
horizontal plane – and exported in some pixel format. When opened in a CAD file the ortho-photo can be
inserted into the existing floor plan and all relevant projection lines can be drawn precisely.
There are some cases which necessitate the acquisition of further data out of the point cloud as, for
instance, the vertical sections of the vaults. As is common usage their profiles are to be retracted around a
horizontal line connecting the two starting points of a vault so that all the vaulting profiles can be added to
the floor plan. How can this information be recovered? Instead of evaluating the point cloud in the way
described above by exploiting a horizontal section, the laser data are cut vertically along a base line which
will define the requested direction – presumably perpendicular to the vault axis – of the intersection. Again
the definition of the thickness of the slice to be selected is vital for the kind of information to be recovered.
The selection of a vertical slice is once again transformed into a polydata and subsequently exported to
CAD.
Fig. 7 – RiSCAN PRO software: single measurements directly taken in the point cloud (Copyright: J. Kanngießer – M. Ganspöck)
317
In other situations, for instance when indicated dimensions have to be inserted into CAD plans, it might be of
great help to take measurements directly in the full point cloud. A function of RiSCAN PRO gives us the
possibility to do this easily. In order to facilitate snapping on respective 3d points, parts of the point cloud can
be selected and hidden. The measuring in 3d space this way constitutes an easy tool for taking every
desired extra measurement. The rest is pure CAD drawing. All temporary results that are imported from
RiSCAN PRO can be organized in layers and evaluated.
Developing the procedures outlined above, it has been possible to evaluate the recorded laser scanning data
with the result of a total of three complete floor plans and four vertical sections of the Castle Hochosterwitz
being produced. While recording of the data took about 250 man-hours, another 470 man-hours were
needed in order to elaborate the final results, displaying a ratio of approximately 1:2 for the time required for
the two main steps, impressively revealing the effectiveness of the proposed strategy. The first complete and
highly precise geometrical documentation of one of the best known and frequently visited cultural heritage
objects in Carinthia has thus been produced.
Summary
This paper has shown that complex architectural surveys can be executed primarily using laser scanning
technologies instead of tachymetry and without any loss of the required exactness of the final result. The set
of plans elaborated for Castle Hochosterwitz had been commissioned to be drawn in the scale of 1:100 but
when it comes to the richness of their detailing the finally executed plans would without any doubt fulfill the
requirements of a 1:50 documentation. Moreover it has been demonstrated that the total time of 730 manhours needed for executing and elaborating a survey of the size of a fairly large castle such as
Hochosterwitz was very low if not considerably lower than compared to a tachymetrical survey.
Fig. 8 – Examples from the final set of plans: ground floor and vertical section (Copyright: J. Kanngießer – M. Ganspöck)
As a further advantage which has been achieved without creating any additional cost, a full surface 3d
documentation of the whole structure has been accomplished; a data set that is complemented by the
recording of all material colours. An extraordinary 1:1 image which constitutes a valuable document of the
status of the cultural heritage structure in the year 2009 has thus been produced. As yet another advantage
318
which, in the course of further elaboration works, will easily be rewarded, that 3d document represents an
archive allowing further planning material – as e.g. a site plan or isoline representations, elevations, orthophotos, and triangulated and coloured 3d models – to be achieved without the inconvenience of turning back
on site.
The speed of the technological progress in scanning hardware and software and in computer technology, in
the future will, without any doubt, strengthen the tendencies sketched in this paper towards a laser scanning
dominated architectural survey. The latest innovations on the laser scanning market already show that
today`s scanners are working much faster and even more precisely while executing much higher numbers of
3d measurements.
79
In addition to that, registration processes with higher grades of automation will once
again shorten the effective time on site. Parallel to that development a significant change in the wishes of
clients will surely occur. The adaption of the visual habits of contractors and end-users of architectural and
archaeological documentations will finally lead to a more common double use of the recorded material:
besides an unchanged use of 2d materials in research, planning and building processes, laser scanning data
will more and more be exploited to achieve further steps in the application chain, as in 3d scenes and films
that are going to be produced for exhibitions and presentations of all kinds.
References
CENTRAL-COMMISSION (1889). Kunst-Topographie des Herzogthums Kärnten. Österreichische Kunst-Topographie, Bd. 1, Vienna, p.
111-119, 255
GRUEBER, P. (1925). Die Burg Hochosterwitz. Klagenfurt: Verlag Kollitsch
GINHART, K. (1927). Die Burg Hochosterwitz in Kärnten. Wien: Verlag Filser
DINKLAGE, K. (1980). Kärnten um 1620. Die Bilder der Khevenhüller-Chronik. Vienna: Edition Tusch. - 3-85063-066-8, esp. p. 199
ULLRICH, A. et al. (2004). Hybrid 3D laser sensor based on a high-performance long-range wide-field-of-view laser scanner and a
calibrated high-resolution digital camera. Orlando: Proceedings of SPIE Vol. 5412, pp. 241-249
JANSA, F et al. (2004). Terrestrial laserscanning and photogrammetry – acquisition techniques complementing one another. ISPRS
Com III WG 7, published in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol.
34, Part XXX
ESSER, G. (2006a). Der Einsatz der Image-Laser-Scanning-Technologie am Beispiel der Bauaufnahme des Propylons des
Orakelheiligtums der Fortuna Primigenia in Palestrina (Latium). 9th International Congress “Cultural Heritage and New Technologies”
Workshop „Archäologie & Computer“, 3rd to 5th Nov. 2004, Vienna: Phoibos Verlag, CD-Publication Deutsche Bibliothek ISBN 3-90123264-8
ESSER, G. (2006b). 3D-Laserscanning am Propylon des antiken Praeneste – Erfahrungen mit einer neuen Messtechnologie aus der
Sicht der Bauforschung. Interdisciplinary Colloquium „Von Handaufmaß bis Hightech II“, 23rd to 26th Febr. 2005 in Cottbus (Germany),
published Mainz am Rhein: Verlag Philipp von Zabern, pp. 43-53
79
See technical specifications of modern laser scanners RIEGL VZ400, FARO Photon and ZOLLER+FRÖHLICH Imager 5006i on their
respective websites.
319
ESSER, G. and Mayer, I. (2008). 3d-geometry and 3d-texture. Documenting early-Christian wall paintings at the Domitilla Catacomb in
Rome. 12th International Congress “Cultural Heritage and New Technologies” Workshop „Archäologie & Computer“, 5th to 7th Nov. 2007,
CD-Publication Deutsche Bibliothek ISBN 978-3-85161-002-4
ZIMMERMANN, N. and ESSER, G. (2008). Showing the Invisible – Documentation and Research on the Roman Domitilla Catacomb
based on Image-Laser-Scanning and 3D-Modelling. 35th Annual Conference of Computer Applications and Quantative Methods in
Archaeology (CAA) “Layers of Perception”, Berlin 2nd to 6th April 2007, published Bonn: Römisch-Germanische Kommission and
Eurasien-Abteilung Berlin des Deutschen Archäologischen Instituts, Kolloquien zur Vor- und Frühgeschichte, Band 10, pp. 58-64
ESSER, G. et al. (2009). Die Domitilla-Katakombe in Rom. Baugeschichtliche Erforschung und Dokumentation eines subterranen
Friedhofs der Antike mittels Photogrammetrie und Laserscanning. Poster-Präsentation, 9th Altbergbau-Kolloquium, 5th to 7th November
2009, Montanuniversität Leoben
320

Laser scanning at Castle Hochosterwitz in Carinthia.

Transcription

Similar documents

Legacy Sell Sheet

Senior Edition - AFNORTH International School

UNIVERSAL LASER BORE SIGHTER–DELUXE KIT

Far more than flowers

44887 CL saddle club

Eastnor Castle tourism leaflet

view our 2016 tourist leaflet pdf

family matters summer 2016 - Heritage Family Credit Union

Dr. Daniel Y. Fung Dermal Laser Centre Yaletown, West Vancouver

4:00pm - Ultrazone

Vienna – City of Many Voices

Full-Text PDF

pdf. (download) - Universität Wien

Issue N o 6

Innovation in Content Capture and Process Management